Lynne 35080149ef x86/tx_float: mark AVX2 functions as AVXSLOW ... Makes Bulldozer prefer AVX functions rather than AVX2, which are 64% slower: AVX: 117653 decicycles in av_tx (fft), 1048535 runs, 41 skips AVX2: 193385 decicycles in av_tx (fft), 1048561 runs, 15 skips The only difference between both is that vgatherdpd is used in the former. We don't want to mark them with the new SLOW_GATHER flag however, since gathers are still faster on Haswell/Zen 2/3 than plain loads.		2022-01-29 03:08:16 +01:00
..
asm.h
bswap.h
cpu.c	avutil/cpu: move slow gather checks below in the function	2021-12-21 17:51:17 -03:00
cpu.h
cpuid.asm
emms.asm
emms.h
fixed_dsp_init.c
fixed_dsp.asm
float_dsp_init.c
float_dsp.asm
imgutils_init.c
imgutils.asm
intmath.h	x86/intmath: add VEX encoded versions of av_clipf() and av_clipd()	2021-11-19 11:21:03 -03:00
intreadwrite.h
lls_init.c	Include attributes.h directly	2021-04-19 14:34:10 +02:00
lls.asm
Makefile	x86/tx_float: do not build tx_float_init.c if x86 assembly is disabled	2022-01-27 02:17:46 +01:00
pixelutils_init.c
pixelutils.asm
pixelutils.h
timer.h
tx_float_init.c	x86/tx_float: mark AVX2 functions as AVXSLOW	2022-01-29 03:08:16 +01:00
tx_float.asm	x86/tx_float: add permute-free FFT versions	2022-01-26 04:13:58 +01:00
w64xmmtest.h
x86inc.asm
x86util.asm